(Worksheets B03 and B04). Poast, however, contains both a petroleum
solvent and an emulsifier
and these adjuvants could influence the rate of dermal absorption. These influences cannot be
quantitatively considered in the exposure assessment. As discussed further in Section 3.4,
however, most dermal exposure scenarios lead to estimates of risk that are very far below a level
of concern and the uncertainty concerning the impact of the adjuvants on dermal absorption rate
has only a minor impact on the risk characterization.
3.2.3.4. Contaminated Water -- Water can be contaminated from runoff, as a result of leaching
from contaminated soil,
from a direct spill, or from unintentional contamination from aerial
applications. For this risk assessment, the two types of estimates made for the concentration of
sethoxydim in ambient water are acute/accidental exposure from an accidental spill and
longer-term exposure to sethoxydim in ambient water that could be associated with the typical
application of this compound to a 10 acre block.
3.2.3.4.1. ACUTE EXPOSURE – Two exposure scenarios are presented for the acute
consumption of contaminated water: an accidental spill into a small pond
and the contamination of
a small stream by runoff.
The accidental spill scenario assumes that a young child consumes contaminated water shortly
after an accidental spill into a small pond. The specifics of this scenarios are given in Worksheet
D05. Because this scenario is based on the assumption that exposure occurs shortly after the
spill, no dissipation or degradation of sethoxydim is considered. This is an extremely conservative
scenario dominated by arbitrary variability. The actual concentrations in the water would depend
heavily on the amount of compound spilled, the size of the water body into which it is spilled, the
time at which water consumption occurs relative
to the time of the spill, and the amount of
contaminated water that is consumed. Based on the spill scenario used in this risk assessment, the
concentration of sethoxydim in a small pond is estimated to range from 0.42 mg/L to 6.8 mg/L
with a central estimate of 2.7 mg/L (Worksheet D05).
As with some other scenarios used in this risk assessment, these exposures are implausibly high
given the current use of sethoxydim by the Forest Service. At the upper limit, this spill scenario
involves 200 gallons (257 liters) of a 9000 mg/L solution. This is equivalent to about 6.8 kg of
sethoxydim [257 liters × 9000 mg/L = 6,813,000 mg
•
6.8 kg] or about 15 lbs. This is about 4
times more sethoxydim than the Forest Service used in all of 1999 – i.e., 3.8 lbs as discussed in
Section 2. Again, this scenario is presented in this risk assessment
in the event that the Forest
Service considers increasing the use of this compound.
The other acute exposure scenario for the consumption of contaminated water involves runoff
into a small stream. No monitoring data have been encountered on the contamination of streams
with sethoxydim after ground or aerial applications of the compound over a wide area.
Consequently, for this component of the exposure assessment, estimates of levels in ambient
water are made based on the GLEAMS model. GLEAMS is a root zone model that can be used
to examine the fate of chemicals in various types of soils under different meteorological and
3-13
hydrogeological conditions (Knisel et al. 1992). As with many environmental fate and transport
models, the input and output files for GLEAMS can be complex. The
general application of the
GLEAMS model to estimating concentrations in ambient water are given in Attachment 2.
For the current risk assessment, the methods detailed in Attachment 2 were used with the
exception that the application site was assumed to consist of a 10 acre square area rather than a
10 acre rights-of-way or 100 acre plot. This adjustment reflects the limited use of sethoxydim by
the Forest Service primarily in relatively small nursery plots. The chemical specific values used in
the GLEAMS modeling are summarized in Table 3-3.
Sethoxydim is rapidly degraded in the environment to a variety of metabolites that are structurally
similar to sethoxydim (Bryceland et al. 1997) and whose toxicity appears to be of the same order
or less than the toxicity of sethoxydim (Section 3.1.9.2). The formation
of metabolites impacts
the GLEAMS modeling in that the relatively rapid degradation rates for sethoxydim itself – i.e.,
the transformation from sethoxydim to one of the sethoxydim metabolites – cannot be used for
estimating exposure to sethoxydim and sethoxydim metabolites. As an alternative, half-times for
the transformation of sethoxydim to carbon dioxide, which represent the complete mineralization
of sethoxydim, are used.
For example, in an anaerobic aquatic environment, sethoxydim is rapidly degraded with a halftime
of less than one day. Most
of the degradation products, however, consist of M1-SO and three
other structurally similar compounds. Over a one year period, only 16.6% to 36.5% of the
original sethoxydim is recovered as CO
2
(Shiotani 1990a). These recovery rates correspond to
0.0049 days
-1
(t
½
=141 days) and 0.0028 days
-1
(t
½
=247 days), identical to the degradation rates in
an aerobic environment. For the risk assessment,
the average of this range, 194 days, is rounded
to two significant digits (i.e., 190 days) and used in the GLEAMS modeling.
The GLEAMS modeling yielded estimates of sethoxydim runoff and percolation that were used to
estimate concentrations in the stream adjacent to a treated plot, as detailed in Section 5.5 of
Attachment 2. The results of the GLEAMS modeling for the small stream are summarized in
Table 3-4. These estimates are expressed as the water contamination rates (WCR) - i.e., the
concentration of the compound in water in units of mg/L normalized for an application rate of
1 lb/acre.
The maximum concentrations of sethoxydim in stream water ranged from 0 to about 500 µg/L
(0.5 ppm) depending on rainfall rates. The typical WCR – i.e., mg/L per lb/acre applied – is
taken as 200 µg/L per lb/acre. This is about the peak concentrations that could be expected at
rainfall rates of about 100 inches per year from sand – i.e., 174 µg/L in Table 3-4 rounded to one
significant digit. The upper limit is taken at 500 µg/L, approximately
the peak concentration from
loam soils at rainfall rates of 250 inches per year – i.e., 491 µg/L in Table 3-4 rounded to one
significant digit. The functional lower limit is taken as 0.020 mg/L per lb/acre applied, about the
peak concentration from sandy soil at an annual rainfall rate of 15 inches per year (see Table 3-4,
19.8 µg/L maximum for sandy soil at an annual rainfall of 15 inches per year). In very arid
3-14